pasquino si ferejohn

ARTICLE

The law of the exception: A typology ofemergency powers

John Ferejohn* and Pasquale Pasquino**

1. Introduction

Constitutions are often designed to check the exercise of power, employingsuch devices as bicameralism, executive veto power, special majorities, and,nowadays, constitutional adjudication. To an extent, these checks reflect akind of distrust of those who wield the authority of the state, at least withrespect to the protection of individual rights, and that distrust is at its greatestwhen it comes to the exercise of executive power. Even British constitutionalarrangements permit judges to interfere with the executive if they can be per-suaded that an administrative action is not authorized by a parliamentarystatute. But insofar as modern constitutional governments are limited in thisway, they may be somewhat disabled in dealing with emergencies. When thepublic safety is seriously threatened, there may be a need for quick and decisiveaction that cannot, perhaps, wait for the deliberate pace of ordinary constitu-tional rule. This, indeed, is a central dilemma of a liberal constitutionalgovernment. The rights and protections it provides and preserves can preventthe government from responding efficiently and energetically to enemies thatwould destroy those rights and, perhaps, even the constitutional order itself.

As a result, modern constitutions often have special provisions for dealingwith emergency situations.1 In cases of an urgent threat to the state or regime,constitutions sometimes permit the delegation of powers to a president, or tosome other constitutional authority, to issue decrees, to censor information,and to suspend legal processes and rights. The purpose for which this specialauthority is granted is fundamentally conservative: it is aimed at resolving thethreat to the system in such a way that the legal/constitutional system isrestored to its previous state. Rights are to be restored, legal processes resumed,

© Oxford University Press and New York University School of Law 2004, 210I.CON, Volume 2, Number 2, 2004, pp. 210–239

*Carolyn S. G. Munro Professor of Political Science and Senior Fellow, The Hoover Institution, at StanfordUniversity and Visiting Professor at New York University School of Law**Research Director, CNRS, France, and Visiting Professor of Politics and Law and New York University

1 See generally EUROPEAN COMMISSION FOR DEMOCRACY THROUGH LAW, EMERGENCY POWERS (Council ofEurope Publishing 1995).

Icon-66.qxd 3/19/04 11:03 AM Page 210

at The T

echnical University of B

ucuresti on January 23, 2011icon.oxfordjournals.org

Dow

nloaded from

http://icon.oxfordjournals.org/

The law of the exception: A typology of emergency powers 211

and ordinary life taken up again. This conservative purpose is reflected inthe fact that the executive is not permitted to use emergency powers to makeany permanent changes in the legal/constitutional system.2 Emergencypowers, exercised in this conservative way, have long been thought to be avital and, perhaps, even an essential component of a liberal constitutional—that is, a rights-protecting—government. They are the key to resolving thedilemma faced by such governments when they are under either external orinternal attack.

2. Models of emergency power

Modern emergency powers are, to a greater or lesser extent, modeled on theancient model of the Roman dictatorship.3 The Roman Constitution wasexceptionally complex and contained a very elaborate system of checks onthe exercise of executive authority.4 Each of the executive offices was, in effect,collegial: there were two consuls who shared the highest executive authority

2 Machiavelli’s defense of Roman dictatorship puts its conservative purpose very well:

As is seen ensued in Rome where in so much passage of time no Dictator did anything thatwas not good for the Republic. . . . A dictator was made for a (limited) time and not in perpe-tuity, and only to remove the cause for which he was created; and his authority extended onlyin being able to decide by himself the ways of meeting that urgent peril, (and) to do thingswithout consultation, and to punish anyone without appeal; but he could do nothing todiminish (the power) of the State, such as would have been the taking away of authority fromthe Senate or the people, to destroy the ancient institutions of the City and the making of newones. So that taking together the short time of the Dictatorship and the limited authority thathe had, and the Roman People uncorrupted, it was impossible that he should exceed his lim-its and harm the City: but from experience it is seen that it (City) always benefited by him.

1 NICCOLO MACHIAVELLI, DISCOURSES ON LIVY ch. 34 (Jon Roland ed., Henry Neville trans. 1675) (1517)(in English), available at http://www.constitution.org/mac/disclivy_.htm.

3 The influence of the Roman model of emergency powers—specifically constitutional dictatorship—on modern constitutional thought is complex but travels through the development of republicanthinking. Particularly important is Polybius’s reconstruction of the Roman Constitution, togetherwith writings of Livy and other historians, that eventually led to the “rediscovery” of Roman con-stitutional practices by Machiavelli, Montesquieu, and Rousseau in developing modern republicanideas. In turn, the drafters of the Weimar Constitution, the German Basic Law of 1919, and of theFrench 1958 Constitution were convinced of the importance of creating a constitutional capacityto cope with urgent threats to the regime or the nation.

4 We may consider executive officials as those who bear the imperium—the power to conductmilitary operations outside of the city, and particularly to kill without any legal process those whoopposed those operations. These included the consuls, proconsuls, praetors, and a few others. Eachof these officials was always accompanied by bodyguards bearing the symbols of imperium—abundle of rods (fasces), which would include an axe if he were outside the city. Inside the city, theaxe is removed and his orders are subject to a right of due process, the provocatio. The dictator,when one was in place, not only possessed the imperium but also possessed that power inside ofRome (domi). The axe was not removed from his fasces when he was within the city.

Icon-66.qxd 3/19/04 11:03 AM 11

at The T



Dow

nloaded from

http://www.constitution.org/mac/disclivy_.htm


212 J. Ferejohn & P. Pasquino

and also commanded the armies.5 There were numerous other subordinateexecutive officials who shared some of these powers as well. The RomanSenate, which was at least as much an executive as a legislative body, couldissue edicts and decrees, and it effectively governed Rome when the consulswere away from the city. But the Senate was usually very large and often inter-nally divided. And from quite early, there were ten tribunes who were respon-sible to the popular assemblies, each of whom had veto power over the actionsof the Senate or the consuls. The effect of this elaborate system of checks wasto maintain the complex system of rights to which Roman citizens had becomeaccustomed, but this complex division of authority produced a governmentthat was unwieldy in times of crisis.6

In cases of emergency the Roman Senate could direct the consuls to appointa dictator for a period of up to six months.7 The dictator was authorized to sus-pend rights and legal processes and to marshal military and other forces todeal with the threat of invasion or insurrection for the purpose of resolving thethreat to the republic.8 When he finished this job he was expected to stepdown, his orders were terminated and their legal effects ended, and the statusquo ante was to be restored. In these respects the purpose of the dictator wasfundamentally conservative. In the same way, modern constitutional emer-gency powers are supposed to be conservative as well. Emergency powers inmodern constitutions are to be employed to deal with temporary situationsand are aimed at restoring the conditions to a state in which the ordinary con-stitutional system of rights and procedures can resume operation. Typically,the holder of emergency powers is not permitted to make law but is restrictedto issuing temporary decrees. And of course, the constitution itself is not to bechanged in such periods.

5 The sharing of executive powers sometimes worked through alternation and sometimes by giv-ing each consul an area where he would rule. But there were ambiguous areas where either con-sul could claim authority and how rule was handled in these areas was worked out throughnegotiation.

6 There is some dispute among historians as to whether the “purpose” of executive collegiality wasto protect rights or if this was merely a side effect of that arrangement. ANDREW LINTOTT, THE

CONSTITUTION OF THE ROMAN REPUBLIC 100–1 (Clarendon Press 1999). Lintott argues, for example,that the reason for collegial executives was as a precaution against the death or unavailability ofone or more of the officials. It is hard to know the purpose of a constitutional arrangement whoseorigins are unrecorded so we remain agnostic about such claims.

7 Originally, the dictator was to have served previously as a consul though with time this restric-tion fell away. He was also restricted to dealing with affairs in Italy, a restriction that was honoreduntil the Second Punic War, after which the dictatorship fell into disuse. See Wilfried Nippel,Emergency Powers in the Roman Republic, in LA THEORIE POLITICO-CONSTITUTIONNELLE DU GOUVERNMENT

D’EXCEPTION 5–23 (Pasquale Paquino & Bernard Manin eds., Les Cahiers du CREA 2000).

8 Technically, the consuls could exercise full imperium only outside the city (militiae). Inside ofRome (domi), the consuls’ orders could be appealed to the courts. The dictator, however, had theimperium inside the city as well and his decisions were not subject to appeal.

Icon-66.qxd 3/19/04 11:03 AM Page 212

at The T



Dow

nloaded from



The Roman dictatorship was rediscovered by Machiavelli and describedby Harrington and Rousseau as a necessary component of a well-functioningrepublic.9 We can see those writers as a kind of bridge between theRoman model of dictatorship and the modern idea of constitutionalemergency powers, which we will call the “neo-Roman model.” Modernconstitutions do not invest special powers in a person outside of governmentbut many vest emergency powers in an elected president. The Constitutionof the Weimar Republic and the 1958 French Constitution are exemplaryin this regard.10 An important distinction between the classical and neo-Roman model is that, while the Roman dictator was chosen from amongmen of special virtues and abilities, in the newer scheme of things the personwho is to wield emergency powers enjoys a kind of popular or democraticmandate.

The situation in the U.S. is complex. The framers were certainly aware of theRoman model: “Every man the least conversant in Roman history, knows howoften that republic was obliged to take refuge in the absolute power of a singleman, under the formidable title of dictator, as well against the intrigues ofambitious individuals who aspired to the tyranny, and the seditions of wholeclasses of the community whose conduct threatened the existence of all gov-ernment, as against the invasions of external enemies who menaced the con-quest and destruction of Rome.”11 Indeed, the experience of the Revolutionitself would have been fresh: George Washington functioned essentially as acongressionally appointed dictator for seven years and then returned to hisfarm. But, by 1787 the framers seemed to think that a unified, energetic, and

9 1 NICCOLO MACHIAVELLI, DISCOURSES ON LIVY ch. 34; JAMES HARRINGTON, THE COMMONWEALTH OF

OCEANA 88 (1656), available at http://www.constitution.org/jh/oceana.htm (“But whereas it isincident to commonwealths, upon emergencies requiring extraordinary speed or secrecy, eitherthrough their natural delays or unnatural haste, to incur equal danger, while holding to the slowpace of their orders, they come not in time to defend themselves from some sudden blow; or break-ing them for the greater speed, they but haste to their own destruction; if the Senate shall at anytime make election of nine knights-extraordinary, to be added to the Council of War, as a junctafor the term of three months, the Council of War with the juncta so added, is for the term of thesame Dictator of Oceana, having power to levy men and money, to make war and peace, as also toenact laws, which shall be good for the space of one year (if they be not sooner repealed by theSenate and the people) and for no longer time, except they be confirmed by the Senate and thepeople. And the whole administration of the commonwealth for the term of the said three monthsshall be in the Dictator, provided that the Dictator shall have no power to do anything that tendsnot to his proper end and institution, but all to the preservation of the commonwealth as itis established, and for the sudden restitution of the same to the natural channel and commoncourse of government. And all acts, orders, decrees, or laws of the Council of War with the junotabeing thus created, shall be signed, ‘DICTATOR OCEANAE.’ ”); JEAN-JACQUES ROUSSEAU, THE

SOCIAL CONTRACT AND DISCOURSES 293 (1762) (G. D. H. Cole trans., J. M. Dent & Sons Ltd. 1973)(“The dictatorship”).

10 See WEIMAR CONST. art. 48; FR. CONST. art. 16.

11 THE FEDERALIST NO. 70, at 391 (Alexander Hamilton) (Clinton Rossiter ed., 1999).

Icon-66.qxd 3/19/04 11:03 AM Page 213

at The T



Dow

nloaded from

http://www.constitution.org/jh/oceana.htm



independent executive would obviate the need for such a special office.12 Thus,the U.S. Constitution provides only a limited grant of explicit emergencypowers in article I, which permits suspensions of habeas corpus in some cir-cumstances (though this is often thought to be a power controlled by Congressrather than the president),13 while article II contains or suggests an array ofimplied powers that may authorize presidential emergency rule (such as thepresident’s powers as commander in chief of the armed forces).14 So,American law on this issue remains an ambiguous example, but it does seemclear that an American president has some constitutionally delegated powersto deal with emergencies.15 If that is so, the U.S. case resembles the situationin the Weimar Republic and France, in that the Constitution explicitlydelegates to the president certain special powers to deal with emergencies.

12 Alexander Hamilton wrote:

The experience of other nations will afford little instruction on this head. As far, however,as it teaches anything, it teaches us not to be enamoured of plurality in the executive. Wehave seen that the Achaeans, on an experiment of two Praetors, were induced to abolishone. The Roman history records many instances of mischiefs to the republic from the dis-sensions between the consuls, and between the military tribunes, who were at times sub-stituted for the consuls. But it gives us no specimens of any peculiar advantages derived tothe state from the circumstance of the plurality of those magistrates. That the dissensionsbetween them were not more frequent or more fatal is matter of astonishment, until weadvert to the singular position in which the republic was almost continually placed, and tothe prudent policy pointed out by the circumstances of the state, and pursued by the con-suls, of making a division of the government between them. The patricians engaged in aperpetual struggle with the plebeians for the preservation of their ancient authorities anddignities; the consuls, who were generally chosen out of the former body, were commonlyunited by the personal interest they had in the defense of the privileges of their order. Inaddition to this motive of union, after the arms of the republic had considerably expandedthe bounds of its empire, it became an established custom with the consuls to divide theadministration between themselves by lot—one of them remaining at Rome to govern thecity and its environs, the other taking the command in the more distant provinces. Thisexpedient must no doubt have had great influence in preventing those collisions and rival-ships which might otherwise have embroiled the peace of the republic.

Id. at 393.

13 U.S. CONST. art. I, § 9, cl. 2.

14 Id. art. II, § 2, cl. 1.

15 The complex ways in which republican, specifically Roman, ideas influenced the structure ofthe American Constitution is a large and controversial topic. There is little doubt, from what theywrote in many places and which authors they cited and used as rhetorical models, that many of theframers and their opponents were directly familiar with Roman histories produced by Livy, Tacitus,and Sallust as well as Polybius and that they would also have learned of the ideas of the separationof powers and checks and balances in republican Rome through Montesquieu’s influentialwritings. Hence, there would have been some knowledge of the Roman dictator and of particulardictators, such as the legendary Cinncinatus and Fabius Maximus, and perhaps the constitutionalframers might have seen a need for dealing with emergencies as a requisite for a stable republic. But,

Icon-66.qxd 3/19/04 11:03 AM Page 214

at The T



Dow

nloaded from



Advanced democracies do not necessarily need to use constitutional powerswhen confronting emergencies. They often prefer to deal with emergenciesthrough ordinary legislation. Such legislation may delegate a great deal ofauthority to the executive and may be enacted for temporary periods. Andthere may be a sense that the legislation is in some ways exceptional. But, how-ever unusual it may be, emergency legislation remains ordinary within theframework of the constitutional system: it is an act of the legislature workingwithin its normal competence. Such legislation is, in the postwar constitu-tional systems, reviewable by the constitutional court (if there is one) and isregulated in exactly the same manner as any other legislative act. For example,in Britain we see the succession of Defense Against Terrorism acts and theUnited States has the PATRIOT Act.16 Each is ordinary though time-limitedlegislation. Many antiterrorist laws have been passed in the same way by theGerman and Italian parliaments in the 1970s and the 1980s.

In fact, not all advanced democracies actually have constitutional provi-sions for the exercise of emergency powers. While the 1958 French Consti-tution has article 16, and the German Basic law has article 115a,17 Britishconstitutional arrangements do not recognize any source of executive author-ity that is not delegated to it by an act of Parliament. At least since thedecline of royal prerogative, Great Britain has only one legal way to deal withemergencies: to enact ordinary laws that delegate special authority to thegovernment or its agencies.

In any case, it is a striking fact that, even in those advanced democracieswhose constitutions contain provisions for emergency powers, these powersare not used. There seem to be two possible reasons for this. First, it may be thatthere have not been emergencies of sufficient magnitude to warrant the invo-cation of emergency powers. It is plausible that elected officials are cautious intriggering the use of exceptional powers and, indeed, that caution is probably

of course, they would also have been acquainted with the perversion of the institution by Sulla andCaesar, and so would have seen the dangers of emergency rule as well. As far as we can tell, the pro-vision for suspending habeas corpus in article I probably owes more to English experience with theHabeas Corpus Acts, which permitted the suspension of the privilege by statute, than to an attemptto implement Roman emergency laws. The president’s commander-in-chief powers seem a morelikely place to look for the republican influence on the structure of emergency powers.

16 Prevention of Terrorism (Temporary Provisions) Act 1974, c. 56 (Eng.); Prevention ofTerrorism (Temporary Provisions) Act 1976, c. 8 (Eng.); Prevention of Terrorism (TemporaryProvisions) Act 1984, c. 54 (Eng.); Prevention of Terrorism (Temporary Provisions) Act 1989,c. 4 (Eng.); Criminal Justice and Public Order Act 1994, c. 33 (Eng.); Criminal Justice (Terrorismand Conspiracy) Act 1998, c. 40 (Eng.); Anti-terrorism, Crime and Security Act 2001, c. 24(Eng.); Uniting and Strengthening America by Providing Appropriate Tools Required to Interceptand Obstruct Terrorism Act of 2001, Pub. L. No. 107-56, 115 Stat. 272 (2001).

17 Compare FR. CONST. art. 16, with GG art. 115a (F.R.G.). This long article was passed by a consti-tutional amendment in 1968. It copes essentially with what Germans call Verteidigungsfall, state ofdefense, having in mind the invasion of the country by nothing less than the Red Army!

Icon-66.qxd 3/19/04 11:03 AM Page 215

at The T



Dow

nloaded from



to be applauded. Perhaps, in view of the historical abuses of such powers (forexample, the extensive use of such powers by Indira Gandhi in the 1970s,18

tyrannical misrule under the guise of emergency rule in various LatinAmerican countries, and, perhaps, the experience of the end of the Weimargovernment), political officials have generally decided that the dangerscarried by the exercise of emergency authority are too great to be used inany but the most dire circumstances. Second, it is possible because of theadvance of state-controlled technology for dealing with disorder, that mostemergencies can be successfully managed by the operation of the ordinarylegal-constitutional system. That is, an emergency that might have requiredthe invocation of emergency powers a century ago, can now be handledeffectively by more or less ordinary policing techniques, beefed up with afew extra powers permitting the detention of suspects without charges, andperhaps suspending their access to lawyers. And these extra powers can besupplied by statutes.

Either of these possibilities may explain why, among the advanced democ-racies, only France has employed such powers in the recent past (in 1961) andthat was to deal with the Algerian situation. The Algerian crisis may well havethreatened the French state in some fundamental way; it certainly had under-mined the previous regime. It may, in that sense, have been analogous to thetraditional triggers of emergency power: invasion, revolution, or catastrophicmilitary defeat. By contrast, Germany, Italy, Spain, Great Britain, and the U.S.all chose to deal with domestic and international terrorism by delegating pow-ers to executives by ordinary legislative means. Terrorists in those states were,on these views, more of a nuisance to public order than a profound threat tothe legal-constitutional order.

One may object, of course, that the countries spoken of here are very stableand entrenched democracies that have little need to invoke extreme constitu-tional measures to protect their regimes. None is plagued by such deep inter-nal rifts as to be at risk of internal insurrection much less civil war. Moreover,the international borders surrounding these countries are stable and uncon-tested and there has been little risk of external invasion. The political distur-bances they face from terrorism may be frightening to many people, but theyoffer no deep threat to the existence of the nation, or the constitution, or evento the political regime; perhaps nothing to warrant testing new constitutionalwaters by investing the executive or the security apparatus with vast and pos-sibly dangerous new powers.

While we accept these points, we think that an explanation for the disuseof constitutional emergency powers must lie, in part, in the development ofa new legal model for dealing with emergencies. A new model of emergencypowers—the legislative model—has evolved over the past half century, at least

18 See Burt Neuborne, Supreme Court of India, 1 INT’L J. CONST. L. (I·CON) 476 (2003).

Icon-66.qxd 3/19/04 11:03 AM Page 216

at The T



Dow

nloaded from



for the advanced or stable democracies.19 The legislative model handlesemergencies by enacting ordinary statutes that delegate special and temporarypowers to the executive. This practice implies that emergency powers are to beunderstood as exceptional to the ordinary operation of the legal system andthat, once the emergency subsides, there will be a return to ordinary legal andpolitical processes. In principle, therefore, legislative emergency powers aretemporary. They are also aimed at restoring the prior legal constitutional sta-tus quo and so, in that sense, are conservative, as is the neo-Roman model.Whoever has had their rights suspended will regain them and normal life willresume. Moreover, because emergency powers are provided in the ordinarylegislative process, that process remains in place to regulate the use of thegranted powers. The legislature is expected to monitor the use of the emer-gency powers, to investigate abuses, to extend these powers if necessary, andperhaps to suspend them if the emergency ends.20

Another feature of the legislative model is that the legislature plays a fun-damental role both in recognizing an emergency and in creating the powersto deal with it. As we shall see, this combination of functions is quite differentfrom the Roman model. The Roman Senate, of course, did play an epistemicrole in recognizing the fact of an emergency, but the powers given to the dicta-tor, and the limits of those powers, were determined constitutionally, prior toany particular emergency, and not by the Senate. In recognizing a state ofemergency, in fact, the Senate effectively suspended its own operations andpermitted the dictator to assume the constitutional duty of resolving the emer-gency. In the legislative model, the epistemic and power-creating functions arecombined.21 This is, therefore, a partial departure from the Roman practice of

19 It is important to note that the use of constitutional emergency powers remains common innewer and more fragile democracies: Latin America, Africa, and southern Asia have seen repeateduses of such constitutional powers. Some of these uses might be judged successful, others less so.But the important empirical observation is that emergency powers remain alive and well in lessstable democracies.

20 In many modern constitutions, the legislature retains a role even in constitutional emergencies.Under article 16 of the French Constitution, the legislature remains in session, and it is able toimpeach the president if it thinks he has exceeded his authority. FR. CONST. art. 16.

21 Indeed, congressional recognition of the need to deal with emergencies was apparent in theearly years of the American republic. For example, in view of doubts as to whether theConstitution provided sufficient authorization to the president, Congress enacted legislation per-mitting the commandeering of the state militias in 1792. Two years later President Washingtonpersonally led these troops into action. Washington’s proclamation, calling for the use of statemilitias to suppress the Whiskey Rebellion, reads as follows:

And whereas, by a law of the United States entitled “An act to provide for calling forth themilitia to execute the laws of the Union, suppress insurrections, and repel invasions,” it isenacted that whenever the laws of the United States shall be opposed or the executionthereof obstructed in any state by combinations too powerful to be suppressed by the ordi-nary course of judicial proceedings or by the powers vested in the marshals by that act, the

Icon-66.qxd 3/19/04 11:03 AM Page 217

at The T



Dow

nloaded from



“heteroinvestiture,” where the party declaring an emergency is completelyseparated from the one that exercises that authority.22

The attractiveness of the legislative model relative to the constitutional (orneo-Roman) model rests on a combination of normative and positive judgments.Constitutional emergency powers are thought to be especially dangerous in thatthey give the executive special and difficult-to-control new powers; to undertakethis danger is warranted only in dire circumstances. And, precisely because thecircumstances are so dire, the executive will be very hard to manage or checkshould he abuse the trust the constitution has placed in him. The legislativemodel permits closer legislative supervision of the executive’s use of legislativelycreated authority, and it provides for the timely ending of that delegation when-ever the legislature thinks the emergency is finished or if the executive hasproven untrustworthy. Thus, legislative delegation seems more flexible. Thisvirtue seems more real in a separation-of-powers system in which the executiveand legislative branches are independent of one another; but even where a par-liamentary system is in place, as in the U.K., the legislature may choose to estab-lish independent commissions to monitor the execution of these new powers.

A second condition for the use of the legislative model is that, in circumstancesof emergency, the legislature is actually willing to enact statutes concedingnew, if temporary, powers to the executive. That is to say, it must be believedgenerally that the “normal” circumstances of jealousy or rivalry between gov-ernmental departments will be overridden during emergencies. At least with

same being notified by an associate justice or the district judge, it shall be lawful for thePresident of the United States to call forth the militia of such state to suppress such combi-nations and to cause the laws to be duly executed. And if the militia of a state, when suchcombinations may happen, shall refuse or be insufficient to suppress the same, it shall belawful for the President, if the legislature of the United States shall not be in session, to callforth and employ such numbers of the militia of any other state or states most convenientthereto as may be necessary; and the use of the militia so to be called forth may be contin-ued, if necessary, until the expiration of thirty days after the commencement of theensuing session; Provided always, that, whenever it may be necessary in the judgment ofthe President to use the military force hereby directed to be called forth, the President shallforthwith, and previous thereto, by proclamation, command such insurgents to disperseand retire peaceably to their respective abodes within a limited time;

And whereas, James Wilson, an associate justice, on the 4th instant, by writing under hishand, did from evidence which had been laid before him notify to me that “in the countiesof Washington and Allegany, in Pennsylvania, laws of the United States are opposed and theexecution thereof obstructed by combinations too powerful to be suppressed by the ordinarycourse of judicial proceedings or by the powers vested in the marshal of that district.” . . .

George Washington’s Proclamation Regarding the Whiskey Rebellion, August 7, 1794, available athttp://www.yale.edu/lawweb/avalon/presiden/proclamations/gwproc03.htm.

22 This paragraph is speaking of the ordinary or paradigmatic use of the legislative model in whichthe legislature creates a new statute authorizing and defining executive action after the emer-gency. There is another case that is common in which the executive bases his actions in emergencyperiods on old emergency statutes that were never repealed.

Icon-66.qxd 3/19/04 11:03 AM Page 218

at The T



Dow

nloaded from

http://www.yale.edu/lawweb/avalon/presiden/proclamations/gwproc03.htm



respect to external emergencies—wars, epidemics, and the like—the forces ofpublic opinion in modern democracies seem to ensure that this condition willhold. It is less clear that internally generated emergencies can be reliably dealtwith in this way, unless the rebellious group is a relatively small minority (as inNorthern Ireland).

Even when the normative and positive conditions for employing the legisla-tive model are fully satisfied, the model poses its own risks. For one thing, thelegislature may be unready or unwilling to act in a timely fashion. Second,even if the legislature is willing to enact emergencies laws, that very actionmay implicate it in the conduct of emergency rule and eliminate a valuablecheck or monitor on the executive. And finally, the laws made to deal with theemergency may become embedded in the normal legal system, essentiallyenacting permanent changes in that system under color of the emergency.Liberties won slowly over long periods of time may be subject to rapid erosionin emergencies and these new restrictions, if they are embedded in law, maynot be rapidly restored if they are restored at all.

Thus, the new model of legislative emergency powers seems to rest on twobeliefs—one negative and one positive. The negative belief is that the tempo-rary nature of emergency legislation will prevent it from corrupting the nor-mal legal system. The use of these powers will take place behind a prophylacticbarrier that will permit the restoration of the ordinary legal system when theemergency has passed. This may not be true, of course. Temporary legislationmay become permanent, as happened in Great Britain with respect to theDefense Against Terrorism Acts and as may well happen in the United States ifthe PATRIOT Act is extended. If this happens, the conservative nature of emer-gency legislation will be lost: a permanent change in the legal system will haveoccurred in circumstances of emergency. This possibility is reminiscent of CarlSchmitt’s constitutional dictator, but now it would be a legislative dictator.23

The positive belief is that because the legislature—the part of governmentclosest to the people—actively delegates authority to the executive, the exerciseof that power is more constrained and legitimate and is even, indeed, amplifiedand made more efficient by the fact that this exercise is supported by the legis-lature and, presumably, by the people.

In important ways, then, the apparently democratic notion of legislativeemergency powers is a significant departure from the classical republicantheory of emergency powers. The republican model envisions a virtuous execu-tor of emergency powers—Cincinnatus called away from his plow—who stepsforward for a sharply defined period of time to defend the institutions of therepublic, and then returns to his farm. In the republican model, the executor is

23 See CARL SCHMITT, DIE DIKTATUR (Duncker & Humblot 1921). Schmitt distinguished betweenkommissarische and souveräne Diktatur. Only the first one has the conservative function typical ofthe Roman magistracy. The latter is tantamount to the “constituent power” and is actually thenegation of true emergency powers.

Icon-66.qxd 3/19/04 11:03 AM Page 219

at The T



Dow

nloaded from



called by others (a senate) to the special position of dictator, which is dormantwithin the constitution, and is automatically dismissed when the emergencyends. He needs no more authority than his own (republican) virtue. He is notsuspected of any other motive than pursuit of public safety; his disinterested-ness and impartiality are assured.24

The new model of legislative emergency powers does not have these advan-tages. Legislative emergency powers rest on the idea that the executive needsan explicit expression of popular support in order to take extreme measures indefense of the government. At least if the executive is not a De Gaulle or aWashington, he has a need for popular support and legitimation and cannotcount solely on public recognition of his republican virtue. In other words, heneeds democratic in addition to republican support. We think that this need forlegitimation is chronic in modern democracies, and that it helps explain theemergence of legislative emergency powers, whereby the legislature provides ameasure of democratic support for the executive’s actions. It is possible, ofcourse, that were any of these democracies to face an emergency of greatmagnitude—or one that deeply divided the country—the executive would beforced, as Lincoln was, to invoke extraordinary constitutional authority andnot to wait on legislative delegation that may be slow in coming. But if parlia-ment can be called upon, if it is in session, or can be summoned quickly, thelogic of enlisting its support is hard for a modern executive to resist.

In the next two sections of the paper we develop a theory for analyzingthe models of emergency power discussed above. First, we will lay out a frame-work for understanding constitutions as systems of norms that must operatein a wide variety of circumstances. Here, the idea is that a constitution isconstructed to deal with a range of more or less normal circumstances andmay produce bad results if it is applied outside of those conditions. So there isa need to recognize exceptional circumstances: wars, invasions, rebellions,and so forth. And given those circumstances, some or all of the operation ofconstitutional norms might be suspended. But to make sense of this we mustgive concrete meaning to the notion of “suspension.” This we do by developingthe notion of “derogation” from a norm, as opposed to abrogation of a norm.We argue in the next section that derogation has the appropriate “conservative”form to serve as a basis of emergency powers that fits with each of our models.What makes a circumstance exceptional? Sometimes there is a special need forspeed or decisional efficiency. Armies need to be created and supplied andmoved rapidly from place to place. Some areas of the country might need to beabandoned, and there may be little time to listen to objections from theresidents. Or, there may be special needs for secrecy so that opponents willbe unable to learn of the nation’s aims or plans. Or, there may be a need to

24 Jon Elster argues that this combination of virtue and disinterest was deeply appealing in theFrench Constituent Assembly. Jon Elster, Authors and Actors: Executive-Legislative Relations in FrenchConstitution-making (2003), available at http://www.yale.edu/coic/Index.htm.

Icon-66.qxd 3/19/04 11:03 AM Page 220

at The T



Dow

nloaded from

http://www.yale.edu/coic/Index.htm



stabilize the constitutional system against the nefarious efforts of its enemies.Each of these needs might be met by suspending rights of speech, assembly,and notice that are normally protected constitutionally.

Next, we consider, in section 4, two kinds of constitutional system: a monis-tic system that insists that the normative ordering remains invariant at all times,and a dualistic system that, essentially, creates two separate constitutional nor-mative orders, each operating in its own circumstance. For a monistic system,emergencies are dealt with from within the normative order. For a dualist sys-tem, there is a regulator that switches between the two orders. This regulatormay have a regular institutional existence, but its metaconstitutional role isseparate from its normal functioning. The Roman Senate’s role in recognizingan emergency was distinct from its normal functioning.

3. Norm and exception: The structure of emergency power

We want to pause briefly to elaborate the conceptual structure of emergencypowers. We start with a conceptual analysis of the notions of “norm” and“exception.” By “norm” we mean, first of all, an empirical regularity in thenatural world or in the society. The norm is more or less strict (it is absolute, insome natural cases, as with the law of gravity, or simply statistical, as when wespeak of the average temperature of New York City in January). By “more orless strict” we mean that the empirical reality might display exceptions, differ-ences, deviations vis-à-vis the norm (for instance, it might exceptionally be55 degrees in New York City in January), or, on the contrary, show no excep-tions (in its field of application, there is no exception to the law of gravity). Insociety we can think of the norm as consisting, say, in arriving on time at ameeting and not leaving before the end of it—it can be remarked that a devia-tion vis-à-vis this norm, if it only happens exceptionally, demands a justifica-tion. The absence of this justification being a deviation vis-à-vis the deviation!

By norm we mean also “a command, a prescription, an order.”25 In public law,the (written or unwritten) constitution represents a set of norms addressing thegovernment or public authorities, which those officials must respect. There is aquestion, which we will not pursue here—asking how it happens that norms, inthe sense that Kelsen gives to this term, are able to become a material source ofobligation? Constitutional norms have specific characters, from a formal andmaterial point of view. Formally, they are norms that regulate and restrict theprinciple of majority rule, which means that, in general (there are few excep-tional cases here also), they can be modified only by special procedures or specialmajorities. Materially, they mostly concern the separation of powers, i.e., thecompetences of the branches of the government and citizens’ rights.

25 HANS KELSEN, GENERAL THEORY OF NORMS 1 (Michael Hartney trans., Clarendon Press 1991).

Icon-66.qxd 3/19/04 11:03 AM Page 221

at The T



Dow

nloaded from



We want to stress that our conceptual analysis concerns the notion ofemergency powers as an exception or derogation vis-à-vis the constitutionalgovernment, meaning here normal or ordinary government. As to derogation,it involves two different issues: the government derogating from its regularform, on the one hand, and the justification of the deviation from the norm, onthe other. We will come back to this later.

Before that, however, a short remark on the word “derogation.” HansKelsen, in his opus posthumous, Die Allgemeine Theorie der Normen, devotes anentire chapter to the notion of derogation/abrogation.26 We are not going tocomment here on Kelsen’s text. It would be interesting, certainly, but it is notour purpose. It is enough to say this: Robert, in his dictionary of the Frenchlanguage, defines “derogation” as the antonym of “conformity, observance”(to a juridical norm, we may add).27 Derogation would be a deviation from nor-mal behavior, allowed under certain circumstances, justified, and in any eventin need of a justification.

Now the sense of the Latin word is somewhat different. Derogare28 isopposed to, and has to be distinguished from, abrogare29 and from obrogare,30

as we can see in a canonical text by Cicero: “huic legi nec obrogari fas est,neque derogari aliquid ex hac licet, neque tota abrogari potest,”31 which maybe translated as “it is not allowed to replace this law with a new one, neither tocancel part of it, nor to abrogate it in its entirety.” In this idea of partial non-application there is something very important concerning the constitutionalderogation, which has an affinity with Carl Schmitt’s distinction betweenVerfassung (the constitution) and Verfassungsgesetze (constitutional norms),where the latter can be suspended in order to protect and stabilize the first.32

So, we have, conceptually:

(1) the norm (in public law: the constitution or, more exactly, the regulargovernment);

(2) the derogation (the emergency powers); and(3) the justification of the derogation.

26 Id. at 106–14.

27 LE PETIT ROBERT, DICTIONNAIRE DE LA LANGUE FRANCAISE 668 (1993).

28 To repeal part of a law, to enforce it only in part. The Oxford English Dictionary gives the samedefinition of the word derogation: “partial abrogation or repeal of a law.” OXFORD ENGLISH DICTIONARY

504 (2d ed. 1989).

29 To repeal a law in its entirety.

30 To abrogate a law replacing it with a new one; to propose a law in opposition to an existing law.

31 CICERO, DE REPUBLICA, DE LEGIBUS III 22 (Clinton Walker Keyes trans., Harvard Univ. Press 1943).

32 This is actually one of the main theses of his chief work. CARL SCHMITT, VERFASSUNGSLEHRE

(Duncker & Humblot 1928).

Icon-66.qxd 3/19/04 11:03 AM Page 222

at The T



Dow

nloaded from



These three elements are connected, moreover, with:

a) the emergency, or, the special circumstances, that must justify the deroga-tion, or the “partial deviation/suspension”; and

b) the higher principle that one may want to shield, protect, or safeguardunder extraordinary circumstances.

Thus, emergency powers exhibit, conceptually as well as normatively, a con-servative aspect. If the exercise of emergency powers undercuts or substan-tially modifies the legal order or the constitution itself it is not only aviolation of norms regulating these powers, it is no longer properly an exerciseof an emergency power at all but is an exercise of constituent power. It isan abrogation or transformation of the constitution and is not functioning topreserve it.33

Of course, such an action is normatively proscribed. In this respect, a con-stitution that has provisions for emergency power already exhibits a dualisticelement: there are two distinct normative systems, insulated from one another:the normal system of rights and procedures and the normative system operat-ing in the state of emergency. This point applies, as well, to the legislative emer-gency model. There is still a distinction between emergency and constitutivepowers, and effecting permanent changes in the legal system under legislativeemergency powers is an exercise in constituent authority. Normally suchaction would be proscribed: in exercising legislative emergency powers, theexecutive is supposed, normatively, to be acting only to resolve the emergencyand restore the normal legal order.

4. Legal dualism and a typology

Now let’s consider the following dichotomy:

(1) Regular government, and(2) Exceptional government.

Some doctrines (we will call them “monist”) claim that there is no differencebetween (1) and (2), either because of the possibility of reducing everything tothe rule of law (the regular government, as is the case with Leibniz andCondorcet) or because of the identity between government as such and the

33 Of course, this is a conceptual analysis and has nothing to say about whether or when theexercise of such constituent powers may be justified. Indeed, it is a common observation thatserious constitutional reforms can only occur in circumstances of emergency. This may imply,depending on who is saying it, a kind of justification of the permanent changes in the U.S.Constitution that were instituted at such times as, for example, the end of the Civil War or duringthe New Deal. We have nothing to say on this point.

Icon-66.qxd 3/19/04 11:03 AM Page 223

at The T



Dow

nloaded from



principle “salus populi suprema—et sola—lex esto” (as with Hobbes and mostof the absolutist doctrines).34

Thomas Hobbes, in chapter 30 of his Leviathan wrote: “Summi imperantisofficia . . . manifeste indicat institutionis finis nimirum salus populi.”35 Here“salus populi” is the norm of the government, not the principle that justifiesa derogation vis-à-vis a regular government. “Salus populi” is, instead, thevery principle leading the ordinary government.36

The dualists (among whom we will include Romans, neo-Romans, orrepublicans) believe normally that (1) is characterized by the existence of indi-vidual rights and by a polyarchy in the structure of the government. Theminimalist form of (1) is the Roman Constitution, as Polybius presents it inthe book VI of his Histories: its polyarchic structure is a form of what theGreeks called “mixed regime or mixed constitution.”37 It is characterized bythe existence of a plurality of magistracies: consuls, senate, tribunes, comitia,and by the provocatio ad populum, the possibility for the Roman citizens toundergo a due process in a case of possible conviction under the death penalty.One sees that in this context the Roman liberty, contrary to what Constantused to think,38 cannot be reduced to the political participation (the so-calledpositive liberty).

Conversely, (2) is represented by the dictatorship that consists in the tem-porary suspension of the right of the provocatio and of the polyarchic structureof the republican government in favor of a monocratic power superior to theindividual rights. We have, therefore, the following dichotomies: polyarchyversus dictatorship, due process versus conviction without trial. Inside the

34 On Leibnitz, see ALFRED BÄUMLER, DAS IRRATIONALITÄTSPROBLEM IN DER ÄSTHETIK UND LOGIK DES 18.JAHRHUNDERTS (1923) (Darmstadt 1981); on Condorcet, see FRANCK ALENGRY, CONDORCET (F. Alcan1904); and on Hobbes’s usage of Salus populi, see THOMAS HOBBES, LEVIATHAN ch. 30 (RichardTuck ed., Cambridge Univ. Press 1991) (in English). Not the doctrines of the raison d’état. On thesedoctrines, see F. SAIT-BONNET, L’ÉTAT D’EXCEPTION 205–24 (Presses Universitaires de France 2001).

35 We quote the Latin version where one finds the formula: “salus populi.” HOBBES, supra note 34,at 231 (“The office of the soveraign . . . consisteth in the end, for which he was trusted with thesoveraign power, namely the procuration of the safety of the people.”).

36 We may remark that in the Hobbesian conceptual universe there is just no room for an emer-gency government since in it norm and exception coincide. The extreme nature of the conflict andthe threat that characterizes his conception of social relations allows this reductio ad unum of ruleand exception. We should bear in mind that Hobbes lived in a society plagued by the overwhelm-ing violence of the religious civil wars. His state of nature is only the abstract and stylized versionof this irreducible type of conflict—and Leviathan (the absolute state) only the political answer tothe state of nature.

37 POLYBIUS, THE RISE OF THE ROMAN EMPIRE 311–17 (Penguin Books 1979).

38 Benjamin Constant, Speech given at the Athénée Royal: The Liberty of the Ancients Compared withthat of the Moderns, in BENJAMIN CONSTANT: POLITICAL WRITINGS 309, 326 (Biancamaria Fontanatrans., and ed., Cambridge Univ. Press 1988).

Icon-66.qxd 3/19/04 11:03 AM Page 224

at The T



Dow

nloaded from



Republican tradition this model39 survived, quite unnoticed (most notably byrecent scholars), through Machiavelli, Harrington, and Rousseau and took aspecial form (the collegial and accountable dictatorship of the Comité de SalutPublic) under the French Revolution. It survived, moreover, as we noticed, inthe public law culture of the European continent in, for example, article 48of the Weimar Constitution (Diktaturgewalt des Reichspraesident) and article 16of the French Constitution of 1958. The characteristic element of the neo-Roman tradition is seen in the fact that it gives a specific place to emergencypowers inside the constitution; they are, so to speak, “constitutionalized.”

There is another model that can be considered dualist. We find it in Englishclassical constitutional theory in connection with the doctrine of the king’sprerogative (praerogativa regia). As it appears in John Locke’s Second Treatise, itseems to contemplate derogation from the regular government based on theseparation of powers and the subordination of the executive power to the leg-islative one. The king’s prerogative is a monocratic power “contra et extralegem” (contrary and beyond the law) in order to protect the salus populi andthe constitutional order.40 It is worth noticing that this doctrine has a demo-cratic or, rather, “monarchomachic” dimension, since the people are the finaljudge of the abuses of prerogative.

It seems that, during the evolution of the English Constitution, the systemmoved from the polyarchic structure (king in Parliament) to a monocratic one(absolute parliamentary sovereignty—where the Parliament is reduced to asingle assembly) so that the U.K. would represent, nowadays, a rare case ofdemocratic/elective form of absolutism (where strictly speaking there is noroom and no necessity for (2)). We saw, nonetheless, that when the BritishParliament passes laws, which derogate the ordinary rights of citizens, asoccurred during the long crisis in Northern Ireland, the Parliament is keen toqualify these statutes as “temporary” acts. True, those laws have been reen-acted repeatedly, after each deadline, for more than twenty years. Even so, ateach occasion the law was time limited. In the end, however, the terrorism lawwas made permanent, showing, perhaps, the difficulty of keeping separate the

39 It is important to stress that we speak here of model also in the sense of an idealization. Thisstarts with Livy and the Annalists and was revived by Machiavelli. As to the historical analysis ofthe dictatorial magistracy during the Roman Republic it would demand a much more nuancedanalysis, which should take into account the works of Theodor Mommsen and ArnaldoMomigliano, André Magdelain and Claude Nicolet. See 2 THEODOR MOMMSEN, ROEMISCHES

STAATSRECHT (Sonderausgabe der wissenschaftlichen Buchgemeinschaft, 9th ed. 1903); ArnaldoMomigliano, Ricerche sulle magistrature romane, reprinted in QUARTO CONTRIBUTO ALLA STORIA DEGLI

STUDI CLASSICI E DEL MONDO ANTICO 273–83 (1969); ANDRÉ MAGDELAIN, JUS IMPERIUM AUCTORITAS: ÉTUDES

DE DROIT ROMAIN 567–88 (École française de Rome 1990); and Claude Nicolet, La dictature à Rome,in DICTATURES ST LÉGITIMITÉ 69–84 (Maurice Duverger ed., 1982).

40 Locke’s prerogative has a much larger scope, actually. It goes from the Aristotelian epieikeia(the doctrine of justice) to the emergency government. See Pasquale Pasquino, Locke on King’sPrerogative, 26 POL. THEORY 198–208 (1998).

Icon-66.qxd 3/19/04 11:03 AM Page 225

at The T



Dow

nloaded from



normal and exceptional legal systems within a legal system based on a principleof parliamentary sovereignty, or more precisely, on a gouvernement d’assemblée.

A dualistic theory of government assumes that “there are exceptional cir-cumstances,” which means circumstances that cannot be governed by regularmeans. We think that it is useful to distinguish between an ontological and anepistemological dimension of the Ausnahmezustand (the exception). From theontological point of view, dualists claim that norm and exception are two differ-ent states of the world. Some (let’s call them “realists”) claim that this onto-logical difference is objective and evident, that everybody can recognize itsexistence or supervenience; as a result, a neutral, involuntary mechanism canbe established in order to detect its appearing or disappearing as a state of theworld. An analogy would be as when the level of a river goes over a giventhreshold and an automatic mechanism systematically triggers an alarmor some other action or performance. Some other thinkers (let’s call them“skeptics,” or we may call them Schmittians) claim that there is no absoluteevidence of the existence of an exceptional situation, that people willinevitably disagree about its existence, and thus we need to attribute to someagency (organ or institution) the epistemic authority to declare the exception.The skeptics believe in the epistemic dimension and are reserved concerningthe ontological one.

From the point of view of political and constitutional theory, the exceptiontakes the form of a special threat to the political order. It is the declared exis-tence of an exceptional threat vis-à-vis the Roman republican system or theconstitutional democratic order of a given political community that triggersand justifies (2)—the emergency regime. The logic of (2) is conservative or pre-servative (in Schmitt’s language we have here a kommissarische Diktatur).41 Thefunction of (2) is to reestablish (1) in its integrity, and as soon as possible.

5. Controlling emergency powers

The Roman dictatorship was in regular use for three hundred years: fromthe establishment of the republic until the defeat of Hannibal. According toNippel, it was used ninety-five times in this period, and its restrictions weregenerally honored.42 No dictator served more than six months, and untilthe war with Hannibal dragged on, none led armies outside of Italy. Part of thereason for this success was that, until the wars with Carthage, the structure ofemergencies facing the Romans fit very well with the institution of the dicta-tor. The fighting season in Italy was roughly from March to late October, so asix-month term was appropriate. Threats sufficient to invoke a dictatorship

41 SCHMITT, supra note 23.

42 Wilfried Nippel, Emergency Powers in the Roman Republic, in LA THEORIE POLITICO-CONSTITUTIONNELLE DU

GOUVERNMENT D’EXCEPTION 5 (Pasquale Pasquino & Bernard Manin eds., Les Cahiers du CREA 2000).

Icon-66.qxd 3/19/04 11:03 AM Page 226

at The T



Dow

nloaded from



generally arose from Rome’s neighbors or from an invasion of the area aroundRome itself.43 And, if the crisis arose from internal turmoil, it could usually beresolved in a short space of time, or at least, the acute aspects of it could be.

The control of the dictatorship was a complex mix of ex ante, interim, andex post mechanism. Ex ante controls were the heteroinvestiture: the fact thatthe entity (the Senate) that declared an emergency requiring a dictator wasseparate from the one exercising choice of the dictator; the prespecified six-month term; the prohibition on leading armies outside of Italy. Interimcontrols were more nebulous. The dictator could, after all, exercise absoluteauthority, and his commands could not be appealed and were subject to nointerference from either the tribunes or the Senate. The Senate, nonetheless,had the control of the budget during the dictatorship. But, in practice, the dic-tators were chosen largely from among those who had held the consulship orperhaps from others with acknowledged military prowess and civic concernand would likely have had no higher ambition than to resolve the crisis facingtheir city. The constitutional norms were, moreover, “bright lines,” so that itwould be clear when a course of action was a constitutional violation. But thismechanism of control depends on the political morality of the person occupy-ing the dictatorship and could not be relied on to control a dictator who failedto respect constitutional morality. Thus, there would be a need for ex postmodes of control of the kind the Greeks would exercise over their military andpolitical leaders. But in Rome these modes were very weak. In principle, a vio-lation might trigger a popular or political reaction against the dictator or hispolicies. In 217 B.C., for example, according to Livy, the Romans rejectedFabius’s policy of pursuing and harassing Hannibal rather than attacking himin force and elected a plebeian consul who promised to be more aggressive.44

However, that popular decision resulted in the disaster at Cannae, and, lateron, Fabius was elected consul two more times. In any case, from the evidenceavailable for this period, dictators did not disappoint the normative expecta-tions of those who appointed them.

There was, however, a kind of constitutional check on dictatorship asopposed to the dictator. The body that declared the emergency (the Senate)and the body that named the dictator (the consuls) were not the same per-son as the one who exercised the dictatorial powers. From a constitutionalviewpoint, therefore, there was a kind of ex ante control of the invocation of the

43 By the time of Hannibal, Rome’s situation had changed materially. Hannibal stayed a very longtime in Italy and the six-month model was not well suited to deal with the threat he posed to Rome.Moreover, it became clear that the solution to this problem required Roman armies to fight outsideof Italy and, increasingly, to stay away from the city for years at a time. The norms of dictatorshipdid not provide for these eventualities and the dictatorship was not again used until the irregularappointments of Sulla and, finally, Julius Caesar.

44 22 LIVY, ROMAN HISTORY 8, 5–6.

Icon-66.qxd 3/19/04 11:03 AM Page 227

at The T



Dow

nloaded from



dictatorship. Sill, once a dictator assumed the office, the only control that couldbe exercised over him involved an appeal to his honor or to his generosity.45

5.1. Modern emergenciesModern circumstances of emergency are very much different from those facedby Rome, and this seems especially true after the events of September 11. Weare faced, nowadays, with serious threats to the public safety that can occuranywhere and that cannot terminate definitively.46 International terrorismrepresents a form of emergency so unlike any Roman circumstance that it isnecessary to reexamine the Roman model to see if it retains lessons for how ademocratic political system should be organized. If we think that the capacityto deal effectively with emergencies is a precondition for republican govern-ment, then it is necessary to ask how emergency powers can be controlled inmodern circumstances.

There are at least two characteristic features of the problem of emergenciesin modern times; first, as mentioned above, contemporary emergencies cannoteasily be limited in time or space. This raises the specter of needing a permanentemergency regime and, in such a circumstance, the Roman practice of eitherbeing in a state of emergency or not may be too rigid. We may need to developan emergency regime that operates alongside the normal regime. That is, it maybe necessary to create legal boundaries around emergencies to substitute for thegeographic and temporal ones that no longer exist. Second, it is no longer clearprior to an emergency what powers are needed to cope with it. The Romanssimply gave the dictator absolute power for a certain period of time to solve theproblem. We may not need to or wish to go that far. We may insist, indeed, thatthe emergency legal system actually be a kind of legal system in which there arerules, rights, and procedures, however limited these may be in its operation.

Modern democratic regimes may demand more interim and ex post controlon the exercise of emergency powers. In American history we can see thisdemand expressed in two ways. The courts have been willing to play, or at leastto consider playing, a somewhat more active role in monitoring the suspensionof rights during emergencies.47 Even if civil libertarians complain that the

45 Livy provides an account of an attempt to prevent a dictator from condemning his second incommand. The supporters of the condemned man appealed to the tribunes to veto the dictator’scommand. After an appeal to the popular assembly by the dictator, none of the tribunes were will-ing to take this step. In the end, the dictator relented as a matter of discretion. 8 LIVY, supra note44, at 33–35.

46 There is another question of how serious these modern emergencies are compared with thosethat threatened Rome, or Weimar Germany, or even the French Fifth Republic in its early years. Itis hard to say that international terrorism poses any serious threat to the existence of a moderndemocratic state. But there are worries that terrorism might escalate technologically to the pointthat the threat could become more profound.

47 Ex parte Vallandigham, 68 U.S. 243 (1863); Ex parte Milligan, 71 U.S. 2 (1866); Ex parteQuirin, 317 U.S. 1 (1942).

Icon-66.qxd 3/19/04 11:03 AM Page 228

at The T



Dow

nloaded from



courts remain too passive in the face of executive assertions of the need forsecrecy, many courts have been open at least to hearing complaints and argu-ments in favor of a judicial role. Even when the courts refuse to intervene,there is usually sufficient dissent from the bench to encourage continuedattempts at litigation.

But the second response is, perhaps, more interesting in the terms of thispaper. The nature of modern emergencies may make a more flexible modelattractive, one in which the appropriate legal instruments can be tailored tothe actual circumstances. This may be a reason for the more extensive use ofthe legislative model. Again, the movement toward the legislative model is notunequivocal, and some writers think that, in many circumstances, the execu-tive has adequate constitutional options. Among modern thinkers who sup-port constitutional emergency powers, we can make out two distinct positions:(1) those who claim that “necessitas non habet legem” (for example, ChiefJustice Rehnquist)48 and (2) those, on the contrary, who claim that there areor should be legal and even constitutional norms that regulate the emergencygovernment. Among the latter we may want to distinguish: (a) the neo-Romans, who claim that exceptional government has to be regulated ex anteby constitutional ad hoc provisions, and (b) those who believe that laws,special laws, or executive measures are better able to confront the crisis.

It seems that the different positions come both from different legal tradi-tions and different normative worries. As for (a), it seems to be supported by“classical republicans” and the exponents of European continental doctrine.Both are particularly aware of the fragility of the constitutional order, becauseof its polyarchic structure, and believe that the constitutional regulation ofexceptional powers can help a nation face not only an emergency but the abuseof emergency powers. The argument goes as follows: if some special type ofgovernment is activated to face an exceptional threat, the attention of thecitizens and of the regular branches of the government are constantly alertand can monitor possible abuses of emergency powers. We can call this the“argument of salience.” How it is that the public can be made to pay attentionto possible abuses is an interesting question of institutional design. Committeesmay be put in place to monitor the use of executive powers or, perhaps, ordi-nary courts could play a role. The British have employed special commissionsto monitor the executive in the Defense against Terrorism legislation.

Those who espouse the second alternative (b) seem less worried about theexercise of emergency powers and its possible abuse by the regular branches ofgovernment. They seem to be more satisfied with ex post control by the judici-ary on the measures taken to face the emergency, and this seems to be thechoice of the American constitutional system. As we saw, a sort of emergencygovernment is spelled out by the U.S. Constitution going under the name of“suspension of habeas corpus,” but this exceptional regime has never been

48 WILLIAM REHNQUIST, ALL THE LAWS BUT ONE: CIVIL LIBERTIES IN WARTIME (Knopf 1998).

Icon-66.qxd 3/19/04 11:03 AM Page 229

at The T



Dow

nloaded from



applied after Abraham Lincoln and the Civil War, and it seems now to beessentially an object for American constitutional history, rather than practice.

There are perhaps no clear criteria for preferring (a) to (b), or vice versa.From the point of view of constitutional engineering much depends on thespecific political context. General recipes are probably not what we wouldsuggest from an armchair. If we consider the dangers inherent to both (a) and(b), we can offer a couple of observations.

5.2. Constitutional mechanisms for controlling emergency powersThere are, speaking from the point of view of historical experience, two kindsof mechanisms with which to check emergency powers (EP).

On the one side, one can use the Roman model of disjunction between theagency declaring the emergency and the agency exercising emergency powers(the Senate and the dictator, or, in a contemporary context, a Constitutional orSupreme Court and the president or the prime minister). And make the agencyexercising EP unaccountable—legally and politically. On the other side, onecan accept that the same agency declares the emergency and exercises EP butmake the decisions taken accountable to courts or some other instance. Thefirst alternative is a mechanism of ex ante control of abuse, designed to ensurethat the agency declaring the emergency has no institutional or private incen-tive to do so. The second is a mechanism of ex post control (somehow exempli-fied by the Supreme Court decisions Milligan and Korematsu).49

Once we have distinguished constitutionalized and nonconstitutionalizedEP, we may want to turn to some other relevant aspects of our topic.

But first, let us sum up what we have been saying in a diagram.

49 Milligan, 71 U.S. 2 (1866); Korematsu v. United States, 323 U.S. 214 (1944).

50 The point here is that when we study and try to classify EP provisions we have to consider fourdimensions: (1) who declares emergency; (2) who exercises these powers; (3) who declares the endof the emergency; and (4) who can interfere with or adjudicate legal questions connected with thedecisions made under EP. For instance, under the Roman republican customary constitution theSenate does (1), the dictator does (2), the constitution stipulates (3) (six months), and nobody does(4) since there are no legal consequences surviving the dictatorship (some citizens were killed bythe dictator but there is nothing to do about that and the dictator is not politically accountable—he does not even exist anymore when his mandate is finished). Think, on the other hand, of thelegislative model. The parliament does (1) and (3). The executive does (2) on the basis of the leg-islative delegation, and the courts do (4).

Constitutional provisions qua declaration (1)concerning EP qua exercise (2)

qua reestablishment of normality (3)qua control on the effects of the emergencymeasures (4)50

Icon-66.qxd 3/19/04 11:03 AM Page 230

at The T



Dow

nloaded from



For example, in the Roman model, according to the scheme above, theSenate controls the declaration of emergency (1), the dictator exercises the EP(2), the constitution itself governs the ending of the EP (3), and there is nobody that can interfere with the exercise of EP (4). We can see that there arefour independent dimensions of EP that permit the classification of constitu-tions in this respect.

Notice that concerning (2) some constitutions (the present German one, forinstance) exclude the suspension of some fundamental rights, while othersexclude the possibility of enacting “laws” (the French president under article16 can only take “measures”).51 Notice, moreover, concerning (3) that withnonconstitutionalized EP, laws can have sunset clauses. And that constitutionsallowing for regulation of EP by other bodies, such as courts or legislatures (4),permit ex post control over the one who exercises EP (2).

5.3. The nature of the threat. Or what to qualify as an emergencyIn the political system we define as a “liberal democratic order,” conflicts are oftwo types: (1) among citizens qua private individuals (private conflicts);and (2) among citizens qua political actors in competition for public offices(constitutional conflicts). The legal system is set up to regulate the first type ofconflict: defining breaches, punishments, and the institutions (the judiciary)in charge of conflict resolution and of retributive justice.

The rules of the political system, instead, regulate the competition forpublic offices—namely, the second type, or constitutional conflict. Emergencymay be defined as a situation that produces a grave disturbance of the politicalsystem or order, threatening its survival. The emergency can have an exoge-nous or/and endogenous origin. The most obvious case of an exogenous threatis a war or invasion: the attempt by “enemies” to destroy, occupy, or somehowtake control over a country (see, for example, article 115 of the GermanConstitution). Special measures are justified in order to protect the integrityof the territory and the very nature of the liberal democratic order. Moreproblematic are two other cases: terrorism and civil war. By civil war we meanthe attempt by internal political actors to destroy the constitutional order(for instance, the Kapp putsch at the beginning of the Weimar Republic; theOAS during the Algerian war in France). Terrorism (internal terrorism, suchas that of the Italian Red Brigades or the German RAF or the French Actiondirecte) seems to be part of the same family. International terrorism may besomewhere between the two—war and civil war.

In these three cases the political order is challenged by enemies who explic-itly want to destroy it, or to modify its nature, so that it seems legitimate to dowhatever is possible in order to render the enemies incapable of representing athreat to the status quo. However, this becomes more problematic when, inorder to prosecute this conflict, our governments suspend our own citizens’

51 See GG art. 81(4) (F.R.G.); FR. CONST. art. 16.

Icon-66.qxd 3/19/04 11:03 AM Page 231

at The T



Dow

nloaded from



rights. Enacting laws introducing crimes like “guilt by association” (see the2001 American PATRIOT Act)52 reminds Italians of the infamous “associ-azione sovversiva” introduced in the 1930s by the fascist criminal code writtenby Alfredo Rocco and adopted by the Stalinist regime and by the authoritariangovernment in Cuba between the two World Wars.53 Secrecy of criminal pro-cedures is also worrisome. One understands pretty well that a governmentstruggling with enemies has good reason to keep secret information that couldhelp the enemy or hinder it, if made public. Nonetheless, one may wonder ifsecrecy is not an instrument to protect the government from the criticism ofinefficaciousness in the very same fight (this is what some tend to believeregarding John Ashcroft’s mania for secrecy—the beauty of it is that becauseof the secrecy one cannot prove such a hypothesis and nobody can disprove it,either). If too much is secret, we are in a very poor position to evaluate or judgeour government. In this kind of situation, we are moving slowly toward adespotic regime (in the Montesquieuian sense of this word). More generally,there is reason to worry that the government may use EP or laws tantamountto EP in order to defeat not enemies but legitimate competitors for office (as theIndian experience under Indira Gandhi shows, in an exemplary way).

A liberal democratic regime can be threatened by a different kind of emer-gency: for example, an economic emergency that, in conjunction with a leg-islative gridlock, triggers urgent and exceptional measures. In this special casethe executive power has to act in the absence of an explicit legislative delega-tion. Ex post approval can be considered, in these circumstances, as the way toreestablish, if possible, the regular pattern of government.

A more acute emergency is that which the Weimar Republic had to face atthe end of its brief life. After 1928 there was no longer in the Reichstag (thelower house of the parliament) any majority support for the constitutionalorder established by the Weimar Constitution of 1919. The liberal democra-tic order inscribed in the Constitution simply had no “democratic,” meaningmajoritarian, support. This tragic conundrum probably has no solution;in other words, there is probably no means, in such a situation, for rescuinga liberal democratic order. Democratically minded citizens could try to engagein a civil war against their enemies. But like in the Lockean “appeal to heaven”this is a challenge and a decision that only each person can take in her/his ownconscience.54

A perhaps less dramatic emergency lies in the unprecedented case ofa threat that seems so dangerous nobody can say how long it will last, or how

52 Uniting and Strengthening America by Providing Appropriate Tools Required to Intercept andObstruct Terrorism Act of 2001, Pub. L. No. 107-56, 115 Stat. 272 (2001).

53 C. p., n. 1398, Oct. 19, 1930, Gazz. Uff., n. 253 (Oct. 28, 1930).

54 JOHN LOCKE, THE SECOND TREATISE OF CIVIL GOVERNMENT AND A LETTER CONCERNING TOLERATION 12, 13,84, 119 (J. W. Gough ed., Blackwell 1946).

Icon-66.qxd 3/19/04 11:03 AM 2

at The T



Dow

nloaded from



it can be dealt with effectively, but is taken seriously enough so that ourfundamental rights are perhaps permanently curtailed. Is internationalterrorism such a threat? We do not know. We will see in the next few years.It seems quite possible that international terrorism will become a more orless permanent part of our future life and will have to be reduced, somehow, toa nonemergency phenomenon. Obviously, this implies somehow a permanentchange in the constitutional system of rights and procedures.

The extreme case is represented by what we may call the possibility of usingemergency powers to institute a “regime change.” To use, in our own way, CarlSchmitt’s language, the enemy is no longer an actor threatening the politicalorder but a political order itself that has become incompatible with the (new)rule one wants to establish.55 Then the emergency is not a conservative form oftemporary government, but the starting point of a new political order. This is largelytantamount to Schmitt’s souveräne Diktatur. One thinks, too, of the Jacobins’project for the regeneration of mankind or of the Leninist dictatorship of theproletariat, which is provisional only in the sense that it represents a transitionto a (brave) new world.

6. Conclusions

There are two main justifications for having constitutional provisions foremergency powers: one is that standard republican institutions suitable forprotecting liberty are too cumbersome to use in emergency situations and sospecial institutions are needed to preserve the republic itself. Machiavelli statedthis justification elegantly in the Discourses:

And truly, among the other Roman institutions, [the dictatorship] isone that merits to be considered and counted among those which werethe cause of the greatness of so great an Empire: For without a similarinstitution, the Cities would have avoided such extraordinary hazardsonly with difficulty; for the customary orders of the Republic move toslowly (no council or Magistrate being able by himself to do anything,but in many cases having to act together) that the assembling together ofopinions takes so much time; and remedies are most dangerouswhen they have to apply to some situation which cannot await time. Andtherefore Republics ought to have a similar method among their institu-tions. And the Venetian Republic (which among modern Republics isexcellent) has reserved authority to a small group (few) of citizens sothat in urgent necessities they can decide on all matters without widerconsultation.56

55 SCHMITT, supra note 23.

56 MACHIAVELLI, supra note 2.

Icon-66.qxd 3/19/04 11:03 AM Page 233

at The T



Dow

nloaded from



But Machiavelli also emphasized that the way that a republic defends itselfought to be regulated constitutionally and not merely improvised.

For when a similar method is lacking in a Republic, either observingthe institutions (strictly) will ruin her, or in order not to ruin her, it willbe necessary to break them. And in a Republic, it should never happenthat it be governed by extraordinary methods. For although the extraor-dinary method would do well at that time, none the less the example doesevil, for if a usage is established of breaking institutions for good objec-tives, then under that pretext they will be broken for evil ones. So that noRepublic will be perfect, unless it has provided for everything with laws[constitutional laws in this context], and provided a remedy for everyincident, and fixed the method of governing it. And therefore concludingI say, that those Republics which in urgent perils do not have resort eitherto a Dictatorship or a similar authority, will always be ruined in graveincidents.57

The second justification for constitutional emergency powers, and the onewe have stressed here, is to protect or insulate the regular operations ofthe legal system from what takes place in emergency circumstances. This isthe argument in favor of constitutional dualism: the notion that there shouldbe provisions for two legal systems, one that operates in normal circumstancesto protect rights and liberties, and another that is suited to dealing with emer-gency circumstances. We think that this second legal system ought to be“legal” in a significant sense and that it ought to be regulated constitutionally.To some extent this regulation may need to rely on something like constitu-tional “morality” or trust of the kind that the Romans evidently had for theinstitution of the dictator, at least for the three hundred years during whichthat institution flourished. But, modern republicans must insist on other formsof protection as well, perhaps by devising special emergency roles for the judi-ciary or the legislature. Indeed, the American experience in wartime—fromthe Civil War cases through Korematsu and Quirin—can be seen as experi-ments in the constitutional regulation of emergency powers.58

But nowadays, we face a somewhat different problem in that nationsare finding other ways to handle emergencies. Even those nations that haveconstitutional emergency powers usually have chosen to employ legislativeemergency powers instead. They choose to handle emergencies, that is, inmuch the same way that nations without such constitutional provisions do.These “dualist” regimes have done this repeatedly despite the apparent alter-ations that such legislative exercises have introduced into their permanentlegal orders. Perhaps, someone will respond, alterations in rights are an

57 Id.

58 Korematsu v. United States, 323 U.S. 214 (1944); Ex parte Quirin, 317 U.S. 1 (1942).

Icon-66.qxd 3/19/04 11:03 AM Page 234

at The T



Dow

nloaded from



inevitable result of technical changes and that such alterations would havehad to occur even if constitutional emergency powers had been employed.Opponents of liberal democracies have so many more vulnerable “soft” targetsand points of potential obstruction and sabotage that the old constitutionalorder has already become anachronistic. Perhaps. But we think that technicalchange is only a partial explanation, and that it is not at all clear that suchchange gives the advantage to the opponents rather than the defenders of lib-eral regimes. We think that the deeper explanation for the resort to legislativeemergency powers is democracy itself: elected officials in every constitutionalsystem need popular legitimacy to engage in the extraordinary legislative proj-ects of the kind required to deal with terrorism. If governments are going tosuspend rights and invade privacy, they need to be able to claim that there iswidespread popular support for such actions. And because those who exerciseemergency powers are elected professional politicians, they are not above sus-picion regarding their private motives. It is easy to employ emergency powersto suppress legitimate opposition and to increase electoral security.

Moreover, there is a constant temptation to permit emergency legislation tospill over into the operation of the ordinary legal system. Rights and libertiesmay be permanently transformed under the threat—real or exaggerated—ofterrorist acts. This is much less possible in the classical regime; constitutionaldictators lack the authority to change the legal system directly. We haveargued that the new dangers arise from the growing demand for democracy ordemocratic legitimation. Perhaps the powers made available in this new modelare more adequate to the task of managing emergencies. Perhaps the older,republican model is no longer even available to us. But we think it worth tryingto envision some possible alterations in modern constitutionalism that wouldpermit a closer approximation to the neo-Roman model.

The critical features of the neo-Roman model are these: first, the aim of theexecutor of emergency powers is conservative. The wielder of these powersis to aim to restore the constitutional status quo ante and not to impose anypermanent change on the legal ordering. In principle, and in fact in some mod-ern constitutions, he is forbidden to do so. He does not possess ordinary leg-islative authority but can issue only ordinances and those ordinances havethe force but not the permanence of laws. Second, the recognition of theemergency is separated from the creation of the specific powers employed todeal with it. The constitutional powers that can be called up during an emer-gency are fixed in advance of the emergency itself. The Roman model itselfagain is exemplary: the powers of the dictator, the length of his tenure, and theplaces that he might lead an army are fixed in advance of his appointment.Finally, the regulation of the emergency powers is fixed constitutionally andoperates before or after the exercise of emergency power itself. In Rome, regu-lation was automatic—in separating the agency declaring the emergency fromthe one invested with the emergency powers (heteroinvestiture), in limiting anemergency to an arbitrary time period, and in various other constitutionalprohibitions. At most, the dictator might be held responsible, politically if not

Icon-66.qxd 3/19/04 11:03 AM Page 235

at The T



Dow

nloaded from



legally, for the conduct of his office after the lapse of the emergency, as whenthe Romans rejected the cautious policy of Fabius Maximus in refusing toengage Hannibal’s army.

Some of these features are available for the legislative model, too. We thinkthat it is normally assumed that the American president’s aims properly are todeal with an emergency without fundamentally altering the constitutionalsystem. If Congress enacts temporary legislation authorizing special presiden-tial powers, the president can then issue administrative orders authorizedby the temporary legislation, and these orders expire with the legislation. He is,thereby, prevented from enacting permanent legislative changes on his own(though he can, of course, issue executive orders). The special danger of thelegislative model is that the authority by which the president takes action isan ordinary statute, and statutes have, intrinsically, the potential to changethe legal system in some permanent way. There is nothing to prevent theformation of an alliance with a congressional majority to make permanentchanges in the law. And, if the justices on the Supreme Court are willing, thispossibility of permanent transformation is even more likely.

Second, within the legislative model, Congress can still play a formal role inrecognizing emergencies, even those that fall short of a constitutionally recog-nized state of war. It has become common for Congress to enact resolutionsrecognizing a state of conflict or justified response—as in the Gulf of TonkinResolution.59 The fact is that such acts do not so much trigger a constitutionaltransition as signal a congressional willingness to further authorize and sup-port presidential action. And, of course, the willingness of Congress to provide(or to refuse) funds for actions taken by the president or to engage in intrusiveoversight are still other modes of regulating and perhaps legitimizing or check-ing emergency rule. Additionally, within the legislative model, Congress canestablish institutional limits on how far the executive may go in employingcongressional authorization, as it has attempted to do in the War PowersAct.60 While no one thinks that this attempt at regulation is fully successful itstands as an example of how the legislative model might develop in the future.It must be realized, however, that if Congress has already recognized an emer-gency and authorized executive action to deal with it, then attempting to tem-per executive actions within the bounds of the legislative model will bepolitically difficult. Furthermore, interfering with an ongoing military opera-tion could expose congressmen, at a minimum, to charges of hypocrisy andirresponsibility. Better perhaps, in this case, to expect regulation from courts.

The courts, after all, can exercise either ex post or continuing control overthe conduct of emergencies. They can enforce the terms of congressionalstatutes in ways that limit the executive’s conduct of emergency operations.They can choose (or refuse) to hear habeas petitions and grant such relief in

59 Gulf of Tonkin Resolution, Joint Resolution of Congress, Pub. L. No. 88-408; 78 Stat. 384 (1964).

60 War Powers Act, 50 U.S.C. § 1541 (1973).

Icon-66.qxd 3/19/04 11:03 AM Page 236

at The T



Dow

nloaded from



wider or narrower sets of circumstances. They can hold officials responsiblefor how they conduct emergency operations under either constitutional orstatutory standards. And, finally, they can refuse to enforce unlawful orunconstitutional orders.

Appendix 1On emergency powers and constitutional architectureWe believe that it is important to insist on the difference between (1) monisticparliamentarism (parliamentary sovereignty) that presently characterizes theU.K. and (2) separation-of-powers systems (Rome and U.S.). But there are somequestions we need to clarify. The world according to Hobbes and the U.K.’s legalorder are both monistic, but for Hobbes it does not make sense to speak of emer-gency power (the salus populi, as we saw, is the only and constant function of thegovernment); conversely, the U.K. Parliament always stressed, explicitly, that theantiterrorist laws were special. It is true that it renewed them for more thantwenty years, but it never made them normal. An emergency can last twenty,thirty, or forty years without ceasing to be perceived as an emergency by the leg-islators. The point here is probably that if there is a prime minister controllingthe parliament, the system is not characterized by “slow motion” [“il mototardo”], as Machiavelli might say. So the system does not need a dictator or apresident with special powers in emergency cases. If there is an American-likesystem of separation of powers, or French (or Weimar) semipresidentialism, thesystem needs a speeding-up mechanism. Sometimes this is explicit and constitu-tionalized (in the neo-Roman regimes); sometimes not, as in the United States,where the Constitution is mixed (like the old English one, but without the lordsand with an implicit prerogative attributed to the president). In a sense, the pre-rogative slowly died in England and moved to the U.S., where it survived in itstamed, republican form. If, in Locke, abuses of the king’s prerogative are checkedby the people’s “appealing to God,” in the democratic American version of theold English Constitution the abuses of presidential prerogative are checked bypublic opinion (polls and elections) and ratified, so to speak, by the SupremeCourt. This is the story told by the democratic popular opposition as comparedwith the pure constitutional engineering model of the Romans.

Appendix 2Note on emergency powers and A. V. Dicey1

“[O]n what principle, and within what limits . . . afford a legal justification foracts . . . which . . . would be breaches of law”2

1 A.V. DICEY, AN INTRODUCTION TO THE STUDY OF THE LAW OF THE CONSTITUTION 396–415 (Macmillan &Co. 1908).

2 Id. at 396–97 (emphasis added).

Icon-66.qxd 3/19/04 11:03 AM Page 237

at The T



Dow

nloaded from



We want to use Dicey’s text to show that a dualistic doctrine of emergencypowers does exist not only in the neo-Roman legal and constitutional traditionbut also in the British constitutional system.

The fact that England became, at the end of the nineteenth century, a quasi-monocratic system based on a unicameral parliamentary sovereignty (whichis clearly an overstatement, since the House of the Lords is still there, withsome small but effective powers) does not prevent its being a dualistic regime,in the sense specified above in Dicey’s text.

For surely the dualism takes, in this context, a special and specific form.First, we are (as in Rome, by the way) under a “customary constitution.”Nothing is actually written down into the constitutional “text,” which—nevertheless—does not mean that the constitution is empty and that wecannot speak of it. Aristotle did, for the Athenian Constitution; Polybius, forthe Roman one. Second, the switch from polyarchy to monocracy thatcharacterizes the neo-Roman dualism is here absent.

But one has to bear in mind that this dichotomy (polyarchy/monocracy) isonly one element of the dualistic regime—at least in our definition.

A second aspect, no less crucial, is the suspension/curtailment of individualfundamental rights. We remember that dictatorship is defined by Livy,Machiavelli, and the classical neo-Romans as suspension of both collegialityand the provocatio ad populum: the Roman liberty.3

Dicey claims that “we must constantly bear in mind the broad andfundamental principle of English law that a British subject must be presumedto possess at all times in England his ordinary common-law rights, and especiallyhis right to personal freedom.”4

Now the question is: How can we justify depriving Englishmen of any oftheir common-law rights, e.g., by establishing a state of martial law—theequivalent of emergency powers?

Here we see that the English legal system is perfectly aware of dual legal-constitutional conditions, I suggest calling these the normal and the excep-tional regimes, the latter being characterized by the provisional/temporarysuspension of fundamental individual rights.

Defining martial law, Dicey insists upon the circumstance that the duty ofthe government is to maintain public order (“the King’s peace”). Now “martiallaw comes to existence in times of invasion or insurrection5 when, where, andin so far as the King’s peace cannot be maintained by ordinary means . . .[because of] urgent and paramount necessity.”6 And Dicey adds, “This power

3 On this point, see 1 FRANCESCO DE MARTINO, STORIA DELLA COSTITUZIONE ROMANA 312, 438 (Jovine,2nd ed. 1972).

4 DICEY, supra note 1, at 397 (emphasis added).

5 Compare the text of the American Constitution concerning the suspension of habeas corpus.

6 DICEY, supra note 1, at 398 (emphasis added).

Icon-66.qxd 3/19/04 11:03 AM Page 238

at The T



Dow

nloaded from



to maintain the peace by the exertion of any amount of force strictly necessaryfor the purpose [principle of proportionality] is sometimes defined as theprerogative of the Crown.”7

Here we see a clear connection between three elements: (1) necessity;(2) martial law (special temporary regime); (3) suspension of individual rights,notably property and personal freedom. “The fact [is] that necessity is the solejustification for martial law or, in other words, for a temporary suspension of theordinary rights of English citizens. . . .”8

Thus we see that the English legal/constitutional system also is aware oftwo regimes: the standard one and martial law. The second triggered by“necessity” is nothing but a temporary suspension of ordinary rights.

That makes martial law similar to the legal condition we qualify inother context as dictatorship, état d’exception, Ausnahmezustand, Notfall,Belagerungszustand. Certainly, we have to take into account the differencesamong these different types of legal dualisms: their (strictly speaking)constitutional status, or, alternatively, their “conventional” status. Moreover,we have to consider and distinguish the organs declaring the emergency(necessity) and exercising the EP. But we think that the universality ofdualism, at least so far as it concerns the nonabsolutist western legal tradition,should be beyond doubt.

7 Id. (emphasis added).

8 Id. at 401 (emphasis added).

Icon-66.qxd 3/19/04 11:03 AM Page 239

at The T



Dow

nloaded from


pasquino si ferejohn

Documents

bernard manin eds

la theorie politico

execution thereof obstructed

provocatio ad populum

york university school

liberal democratic order

ordinary legal system

german basic law